Main

Approximately 2% of all individuals born in the United States today will develop melanoma. Of these, 16% will present with regional or distant metastasis. Unfortunately, the 5-year survival rate for patients at these stages is 62.4 and 16%, respectively.1 The incidence of malignant melanoma has steadily risen over the past few decades. Early metastasis along with a propensity for afflicting a younger population causes a substantial number of years of potential life lost. Melanoma has also notoriously been refractory to chemotherapy regimens, with dacarbazine being the only FDA-approved chemotherapy for the treatment of malignant melanoma. The arrival of BRAF inhibitors, namely vemurafenib and dabrafenib, transformed the way in which patients with malignant melanoma were treated. BRAF inhibitors demonstrate increased affinity for mutated BRAF enzymes over wild type and in this manner they selectively inhibit mutant melanocytes. BRAF mutations are found in ∼50% of malignant melanomas, especially those arising on non-chronically sun-damaged skin.2, 3, 4 When treated with BRAF inhibitors, patients have a significant improvement in progression-free survival as well as overall survival.5, 6, 7 In phase III trials comparing vemurafenib with dacarbazine therapy, vemurafenib-treated patients had a 48% response rate compared with only a 5% response rate in the dacarbazine-treated patients. In addition, there was a 63% relative reduction in the risk of death.6 Therefore, correctly identifying those patients who may derive some benefit through BRAF inhibitor therapy is paramount.

Many methods only detect V600E mutations of the BRAF gene. Newer detection kits identify several less common V600 variant mutations, rather than only identifying V600E. Moreover, there is evidence that many less frequent BRAF mutations are sensitive to BRAF inhibitor or MEK inhibitor therapy.5, 7, 8, 9, 10, 11 Currently those patients with unresectable stage III and IV melanoma are suggested to undergo BRAF mutation analysis and potential molecularly targeted therapy. The analysis, however, typically only takes place on one tissue block from one lesion with the assumption that all lesions from that patient will display similar BRAF status findings. The overall extent and impact of alternate BRAF mutations and discordancy is largely unknown.12, 13, 14 If a significant proportion of patients demonstrate intertumoral discordancy, they may benefit from BRAF inhibitor therapy, which would have not been offered to them initially. In addition, the detection of other less common BRAF mutations would also increase the number of patients offered a BRAF and/or MEK inhibitor. To this end, we studied 25 patients with both primary and metastatic diseases to determine the intertumoral discordance and detection of less frequent BRAF V600 mutations.

Materials and methods

Tissue samples from 177 patients with unresectable and/or metastatic melanoma were sent for molecular BRAF testing between May 2011 and May 2013. Of these, 25 patients with tumor tissue from both primary and metastatic sites were available for further testing. A patient chart review revealed no patients presented with multiple primary melanomas. Patients were 54% female and 46% male with an age ranging from 23 to 81 years. Tissue tested included primary skin, skin metastasis, lymph node metastasis, and one metastatic lesion to the larynx (Table 1). Genomic DNA was extracted using the Qiagen FFPE DNA extraction kit from six sections of formalin-fixed paraffin-embedded material (5 mM thickness each) from a total of 58 tumor samples in formalin-fixed paraffin-embedded tissue. To increase tumor cell population, hematoxylin and eosin slides were microscopically evaluated and marked by a pathologist. Marked areas containing tumor were subsequently microdissected. The area of the slides was at least 0.5 cm2 in general. The limit of detection was 2% for the assays. The DNA concentration was determined using the NanoDrop Spectrophotometer then placed on the BRAF assay. The assay detects five somatic mutations in the BRAF gene using real-time PCR on the Rotor-Gene Q 5plex HRM instrument. The assay uses ARMS (Amplification Refractory Mutation System) and Scorpions technologies; enabling detection of the following mutations at codon 600 of the BRAF oncogene against a background of wild-type genomic DNA: V600E/V600E complex (V600Ec), V600D, V600K, and V600R. The control assay, labeled with FAM, is used to assess the total amplifiable BRAF DNA in a sample. The control assay amplifies a region of exon 3 of the BRAF gene. The primers and Scorpion probe have been designed to amplify independently of any known BRAF polymorphisms. All procedures were performed according to the manufacturer’s protocol (Qiagen Manchester, Manchester, UK) (Figure 1).

Table 1 Comparative data for BRAF status evaluation in primary and metastatic sitesa
Full size table
Figure 1
figure 1

Qiagen results and histologic findings. (a) Raw fluorescence of a mutation-negative sample. Sample control (purple) in the FAM channel true amplification within the acceptable specified range. No amplification present in the FAM channel for any of the mutations. (b) Raw fluorescence of a V600E/Ec sample. Sample control (purple) in the FAM channel true amplification within the acceptable specified range. Amplification for the V600E/Ec mutation acceptable (blue) and the calculated delta CT values lower than the cutoff value for V600E/Ec. (c) Low-power view of V600E metastatic melanoma. (d) High-power view of a V600E metastatic melanoma.

Full size image

Furthermore, a patient chart review revealed that no patients were treated with BRAF inhibitor therapy between primary and metastatic lesion resection. There was one patient for whom clinical information was not available; however, this patient did not demonstrate discordancy (Table 1). This research was approved by the Institutional Review Board.

Results

We analyzed the BRAF mutations in both primary and metastatic sites from 25 melanoma patients. Patients were 54% female and 46% male with an age ranging from 23 to 81 years. Tissue tested included primary skin, skin metastasis and lymph node metastasis, and one metastatic lesion to the larynx.

BRAF status and anatomic tumor location, along with patient demographics, were compiled and are included in Table 1. BRAF status was determined in all 58 specimens from 25 patients in which 37 (64%) were positive for a BRAF mutation (Table 1). Four of the 25 (16%) patients had BRAF mutation discordancy between the primary and metastatic lesions. Of the discordant cases, two (8%) had BRAF V600 mutation-positive primary melanomas with wild-type metastatic lesions and two (8%) patients had a BRAF V600 wild-type primary melanoma with a V600 mutation-positive metastatic lesion (Table 2). Tissue from multiple metastatic sites was available for testing in six patients. A much higher discordancy rate of two out of the six (33%) was found when multiple sites were tested (Table 3).

Table 2 Variable occurrence in discordancy of BRAF mutationsa
Full size table
Table 3 Rate of discordancy in multisite testinga
Full size table

Discussion

Early studies on BRAF heterogeneity proposed that testing of either the primary or the metastatic lesion was sufficient as little discordance existed between different tumor sites.15, 16 More recently, BRAF status discrepancies between primary and metastatic sites range from 18 to 26%.13 We also identify the frequent presence of BRAF discordancy between primary and metastatic melanoma lesions of 16% (Table 1). In addition, we find variability in discordance, that is, a wild-type primary tumor with a mutant metastasis and vice versa (Table 2). Moreover, we show that testing multiple sites in a single patient increases the likelihood of finding BRAF discrepancies (Table 3). Improved patient survival rates with BRAF inhibitors suggest that every opportunity should be taken to detect these mutations when they are present. One major limitation of the current study includes a relatively small number of patients who qualified (multiple tumors, tissue available for testing, etc) even after a broad search was performed. Also, there is a real possibility that other mutations outside the ones identified by the Qiagen test could have been present and not identified as only the most common mutations would be identified (<1%).

Not surprisingly, patients with BRAF V600 mutation-negative melanoma do not derive any benefit from BRAF inhibitor therapy.17, 18 In fact, paradoxically, BRAF V600 mutation-negative tumor cells treated with selective BRAF inhibitors exhibited an increase rate of proliferation, reduced cell adherence, and increased mobility of cells.19 The intertumoral and intratumoral heterogeneity may partially account for the range in initial patient response as well as timing and manifestations of resistance. For maximum patient benefit as well as proper allocation of resources, patients need to be properly classified; including those with less frequent BRAF mutations. The second most common mutation is the BRAF V600K mutation, representing 14% to 28% of BRAF mutations in melanoma.20 We discovered two V600K mutations, which were detected only after using the Qiagen kit, as opposed to the older detection methods (Table 1). This is especially important, as there is emerging literature that many less common BRAF mutations are also responsive to inhibitor therapy including V600K, V600R, K601E, and L597S.5, 8, 9, 10, 11, 21, 22 In our opinion, through optimizing the identification of different types of BRAF mutations, those patients with less frequent BRAF mutations may have access to this life-extending class of drugs, who would have otherwise been excluded from selective BRAF or MEK inhibitor therapy. Additionally, development of therapeutic agents with varied spectrums of BRAF (and MAPK pathway) inhibition may lead to more therapeutic options for patients with non-V600E BRAF aberrations.23 To this end, more updated and advanced detection methods should be encouraged.

Discordancy is frequently present between primary and metastatic lesions of patients with melanoma. Multiple mechanisms may be responsible for these discrepancies (Figure 2). We demonstrate that testing multiple metastatic sites increased the percentage of discordant cases as well as BRAF mutations overall. The clinical and prognostic significance of discordancy is currently unknown. Recent studies evaluating immunohistochemical stains for V600E-mutated BRAF reveals a sensitivity and specificity of 97 and 100%, respectively. Immunohistochemistry was, however, less sensitive in detecting V600 mutations overall (91%).12, 24 If immunohistochemistry proves to be a valid method of BRAF detection, standardization will provide more interobserver uniformity. This may allow for easier assessment of intra- as well as intertumoral heterogeneity and discordancy, respectively. There are conflicting findings among the few papers specifically addressing BRAF heterogeneity in melanoma. One recent paper found no intertumoral heterogeneity between primary and metastatic lesions.25 Whereas these findings are at odds with our results, multiple other studies are in agreement with our reported findings.12, 14, 26, 27 The main difference between these previous studies and our present findings is that we have identified discordancy from both mutated and wild-type primary lesions.

Figure 2
figure 2

Mechanisms of discordant results in BRAF mutation analysis. (a) Intratumoral heterogeneity may show a wild-type primary analysis, yet the metastatic lesions may arise from the mutated cells. The reverse findings may also be seen (ie, a mutated primary analysis and wild-type metastasis). (b) Two separate primary melanomas may yield discordant results if the first primary is tested while the second gives rise to metastasis. (c) Melanoma cells may acquire BRAF mutations over time, evident in metastatic lesions but not present in the primary. At the time of primary excision, already metastatic melanocytes may acquire BRAF mutations after analysis of the primary lesion.

Full size image

Current treatments and strategies only focus on interpatient heterogeneity but there is a growing understanding that perhaps a focus on intratumor and intertumor variability within the same patient should be the new approach.28 Current strategies do not account for evolution of tumors over time, which may have a role in the discrepancy in BRAF in melanoma, and possibly in other genes in melanoma and other cancers. Previous studies from our group demonstrated that significant genetic divergence occurs frequently during the clonal evolution of melanoma and that multiple coexisting metastases appear to be derived from different, genetically unrelated tumor clones.27 Perhaps the evaluation and treatment of melanoma should broaden to accommodate the theory of tumoral heterogeneity and that tumors may evolve over time. In other tumors such as breast carcinoma, the continual monitoring of estrogen, progesterone, and HER2/neu receptors over time is standard of care. In lymphoproliferative disorders it is not uncommon to assess progression and therapeutic strategy based on molecular and immunohistochemical evolution over time. The evidence of discrepant BRAF mutation findings may account for the large variability in response among patients with BRAF inhibitor-treated melanoma. Future studies evaluating the optimal therapeutic strategy for patients with disparate BRAF mutational status are necessary.